The study fulfills objectives of PA 95-070, NIMH, "Brief Interventions to Prevent The Spread of AIDS", by testing an intervention to reduce unsafe sexual behaviors and nondisclosure to partners in HIV-positive persons. This approach to prevention intervenes directly on sources of new infections and attempts to reach sexual partners who might not be accessible through other means. The clinic-based intervention, called "Partnership For Health", emphasizes the importance of a patient-provider team approach in promoting the patient's health and well-being. It integrates preventive efforts with the routine clinical care of HIV patients and does not require substantial resources beyond those already available at the clinics. Unlike traditional posttest counseling, the intervention is given to patients across time and attempts to instill self-protective as well as partner-protective motivation for reducing risk behaviors. Further, the study provides the first theoretical test of the impact of gain-versus loss-framed messages on sexual behavior and disclosure. Six HIV clinics in California will participate. Two clinics implement gain-framed messages emphasizing the positive consequences of engaging in safer sex; two clinics implement loss-framed messages emphasizing the negative consequences of not engaging in safer sex; and two clinics implement an attention-control protocol (focused on diet and exercise). All patients attending the clinics during a nine-month period receive the interventions, delivered through professionally printed pamphlets, communication from medical providers during examinations, and goal setter. Over 6,500 HIV-positive individuals will receive behavioral intervention. It is hypothesized that the gain-framed approach is more efficacious than the loss-framed approach in promoting sexual risk reduction in seropositive persons, and that each approach will surpass the attention-control group. It is also anticipated that the intervention effects will be stronger among individuals with positive attitudes toward condoms and high levels of self-efficacy to enact safer sexual behaviors. The study employs a pre-post design in which a cohort of patients randomly selected at each clinic (n=100 per clinic) is assessed before and after the nine-month intervention period. To disconnect the measurement and intervention components, self-report measures - are obtained under the guise of a general health survey conducted by external researchers.